MISR: Morphology-Integrated Structural Refinement for Reference-Free Underwater Image Enhancement
Underwater images often exhibit color casts, low contrast, and weakened structural details due to wavelength-dependent attenuation and scattering. This paper proposes MISR, a reference-free enhancement framework that combines spatial-domain color and contrast processing with transform-domain structural refinement, without requiring task-specific underwater training. Adaptive Color Consistency (ACC) incorporates a pretrained RGB-uv-based nonlinear color-correction model to perform image-adaptive color correction through feature matching and weighted polynomial mapping. Morphological Contrast Enhancement (MCE) employs top-hat and bottom-hat operations with Contrast Improvement Ratio (CIR)-based adaptive scale selection to enhance local contrast. Structural Response Enhancement (SRE) utilizes the nonsubsampled contourlet transform (NSCT) and normalized local energy to construct adaptive gain maps for selectively modulating directional coefficients according to their relative local energy within individual subbands. MISR is evaluated against ten recent underwater image enhancement methods published within the past three years on four real-world underwater datasets using five no-reference quality metrics and image variance (VAR) as a supplementary descriptive statistic. In addition, paired datasets from UIEBD and LSUI are evaluated using PSNR and SSIM.
Authors
- Xingcheng Han
- Ling Wang
- Bijie Li (ORCID: https://orcid.org/0009-0004-2714-9320)
Institutions
- North University of China (CN)
Publication Details
- Journal
- Journal of Marine Science and Engineering
- Published
- 2026-11-18
- DOI
- https://doi.org/10.3390/jmse14222195
- Primary Topic
- Image Enhancement Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Natural Science Foundation of China